From mboxrd@z Thu Jan 1 00:00:00 1970 From: Tejun Heo Subject: Re: Nvidia ION (MCP79) eSata fails for external cabinet Thecus N2050 (SIL5744) Date: Mon, 09 Nov 2009 17:10:23 +0900 Message-ID: <4AF7CE6F.9050306@kernel.org> References: Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: Received: from hera.kernel.org ([140.211.167.34]:34833 "EHLO hera.kernel.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753844AbZKIIK1 (ORCPT ); Mon, 9 Nov 2009 03:10:27 -0500 In-Reply-To: Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: =?ISO-8859-1?Q?=D6rjan_Nordlund?= Cc: linux-ide@vger.kernel.org Hello, =D6rjan Nordlund wrote: > I am having an issue using eSata to connect my Thecus N2050 storage > cabinet to my new homeserver running Ubuntu Server 9.04. > dmesg shows "failed to IDENTIFY (I/O error, err_mask=3D0x100)" >=20 > I have read forums and mailing lists to understand solutions to simil= ar > problems and the following steps have been taken: > Upgraded to Vanilla kernel 2.6.31 > Changed BIOS settings from SATA->AHCI and then back again > Tried boot with pci=3Dnomsi > Changed cables (i have three cables now and no one works) > Checked that no jumpers can be changed on disks or mainboard > Checked that the disks can be read through USB interface and from oth= er > computers (works on Linux, OS X and Win XP) > Checked disks through eSata on Win XP box > Upgraded BIOS to latest release (August 2009) >=20 > I have looked in all forums and LKML archives for a solution but can = not > really find anyone that has a similar problem. >=20 > I am suspecting the quirks are in the Thecus N2050. I have had some > problems with this device in linux before when i was using USB. Thank= s > to Alan Stern a patch (as1154) was introduced in October 2008 to help > get the device recognized properly by scsi when probing it. >=20 > My Nvidia ION board has a MCP79 controller on-board (should be OK fro= m > 2.6.24) and the Thecus has a Silicon Image controller (i opened the b= ox > and read on the chip) with markings: > SIL5744CNU SIL5744 should work fine. Hmmm... > dmesg (extract from hotplug of device): > [ 753.849929] ata3: exception Emask 0x10 SAct 0x0 SErr 0x1810000 act= ion > 0xe frozen > [ 753.849970] ata3: irq_stat 0x00400000, PHY RDY changed > [ 753.849994] ata3: SError: { PHYRdyChg LinkSeq TrStaTrns } > [ 753.850024] ata3: hard resetting link > [ 754.598415] ata3: SATA link down (SStatus 10 SControl 300) > [ 754.598430] ata3: EH complete > [ 758.663954] ata3: exception Emask 0x10 SAct 0x0 SErr 0x4140000 act= ion > 0xe frozen > [ 758.663993] ata3: irq_stat 0x00000040, connection status changed > [ 758.664018] ata3: SError: { CommWake Dispar DevExch } > [ 758.664046] ata3: hard resetting link > [ 759.588208] ata3: SATA link up 3.0 Gbps (SStatus 123 SControl 300) > [ 759.588245] ata3.00: failed to IDENTIFY (I/O error, err_mask=3D0x1= 00) Looks like the device is failing IDENTIFYs immediately. I don't think this has anything to do with being _e_SATA. Can you try it with other controllers? Thanks. --=20 tejun